Performance analysis of IPv6 transition supporting solutions

OData support
Dr. Jeney Gábor
Department of Networked Systems and Services

The Internet Protocol version 4, which belongs to the Network Layer of the OSI model, does not always support the newest technical requirements. Furthermore its address range is smaller than the number of people living on Earth. Among others this is the reason why a change was necessary to the new Internet Protocol, IPv6. IPv6 has been provided with large address range, mobility support, bigger security and other useful functions in regards of modern technology. The simplest migration would be a single change to IPv6-only mode on all devices of the world at the same time. But it is not even remotely possible since devices must be gradually prepared for the change. This way IPv6 “islands” are formed on the IPv4 “sea” of the world. In order to be able to connect those islands, new mechanisms had to have been developed, and said mechanisms provide IPv6 connection over IPv4 network. These mechanisms are named tunnelling mechanisms. The point of these mechanisms is that the IPv6 packets are embedded into an IPv4 packet, so an IPv6 packet is labelled with an IPv4 header. Due to this embedding, the IPv6 packet can be sent over the IPv4 Internet. There are numerous tunnelling mechanisms.

In the next chapters I would like to discuss an introduction of mechanisms named ISATAP, 6to4 and 6in4. With the help of ISATAP and 6to4 address autoconfiguration mechanisms we get a globally routable IPv6 address. Hosts assign IPv6 addresses to the tunnelling interface from their public IPv4 address. The difference between those two mechanisms is that ISATAP provides IPv6 connection between dual-stack hosts through an IPv4-Intranet, while 6to4 between IPv6 sites and hosts through the IPv4 Internet. However 6in4 requires manual configuration. We need to execute commands to get a global IPv6 address. With these mechanisms we are able to create an almost as good IPv6 connection, as via native connection. Almost, because there is some delay due to embedding and unpacking packets on the border of IPv6 and IPv4 networks.

My thesis is about testing differences between native IPv6 and 6in4 IPv6 connections. Differences are confirmed by measurements. I performed the measurements both on TCP and UDP traffic and drew conclusions from those results.


Please sign in to download the files of this thesis.